Makespan reduction for Time-Weighted Systems using a Clonal Selection algorithm

The time-optimal control of Time-Weighted Systems (TWS) is an approach that can be used to makespan reduction of Computer-Integrated Manufacturing (CIM) Systems. Unfortunately, the complexity of the TWS algorithms turns even suboptimal solutions computationally intractable when dealing with realisti...

Full description

Saved in:

Bibliographic Details
Published in:	2022 IEEE 27th International Conference on Emerging Technologies and Factory Automation (ETFA) pp. 1 - 8
Main Authors:	Vargas, Gabriel Laport, Da Cunha, Antonio Eduardo Carrilho
Format:	Conference Proceeding
Language:	English
Published:	IEEE 06-09-2022
Subjects:	Common Information Model (computing) Computational modeling Control systems Costs Discrete event systems Heuristic algorithms Makespan reduction Manufacturing Production Supervisory control Time-weighted systems
Online Access:	Get full text
Tags:	Add Tag No Tags, Be the first to tag this record!

Description
Summary:	The time-optimal control of Time-Weighted Systems (TWS) is an approach that can be used to makespan reduction of Computer-Integrated Manufacturing (CIM) Systems. Unfortunately, the complexity of the TWS algorithms turns even suboptimal solutions computationally intractable when dealing with realistic CIM systems with multiple components, stations, and different productions batches. In this paper, we propose some strategies to alleviate the complexity cost of makespan reduction computations using TWS. The strategies consist of: relaxing the need of symmetric imposed mutual exclusions relations between events; using a hashtable algorithm for makespan computations instead of a heap of pieces algorithm; avoiding computation of batch supervisors using a scalable synthesis method; and applying a clonal search heuristic for makespan reduction. With this, we obtain the reduction of the computational cost while maintaining the advantages of the TWS approach regarding the reduction of models by representing one task by only one event. The performance of the proposed algorithm is compared with some of the known efficient approaches by means of a realistic CIM example.
DOI:	10.1109/ETFA52439.2022.9921537